at SciVerse ScienceDirect
Journal of Archaeological Science 38 (2011) 3483e3496Contents lists availableJournal of Archaeological Science
journal homepage: http : / /www.elsevier .com/locate/ jasSourcing ceramics with portable XRF spectrometers? A comparison with INAA
using Mimbres pottery from the American Southwest
Robert J. Speakman a,*, Nicole C. Little a, Darrell Creel b, Myles R. Miller c, Javier G. Iñañez a,d
aMuseum Conservation Institute, Smithsonian Institution, Suitland, MD 20746, USA
b Texas Archaeological Research Laboratory, University of Texas, Austin, TX 78712, USA
cGeo-Marine Inc., El Paso, TX 77912, USA
dDept. de Prehistòria, Història Antiga i Arqueologia, Universitat de Barcelona, Barcelona 08001, Catalonia, Spaina r t i c l e i n f o
Article history:
Received 3 May 2011
Received in revised form
11 August 2011
Accepted 15 August 2011
Keywords:
Portable and handheld XRF
INAA
mXRF
Mimbres and Jornada pottery* Corresponding author. Tel.: þ1 301 238 1242; fax
E-mail address: Speakmanj@si.edu (R.J. Speakman
0305-4403/$ e see front matter Published by Elsevie
doi:10.1016/j.jas.2011.08.011a b s t r a c t
Seventy-five intact Mimbres and Jornada pottery sherds from the American Southwest were analyzed by
portable XRF and instrumental neutron activation analysis (INAA). Examination of the data demonstrates
that INAA and portable XRF results for elements common to both analyses can be used to construct
similar compositional groups. When individual compositional groups are compared to one another, it is
apparent that unambiguous separation of compositional groups is challenging by portable XRF given (1)
the limited number of key discriminating elements that can be measured relative to INAA, and (2) the
relative analytical precision and accuracy of portable XRF for measurements of intact heterogeneous
ceramics. We conclude that sourcing intact ceramics by portable XRF is challenging and that bulk
analytical measurements, such as INAA, remain a better approach for sourcing archaeological pottery.
Published by Elsevier Ltd.1. Introduction
Since the first archaeological applications of X-ray fluorescence
(XRF) spectrometry in the 1960s, XRF has emerged as one of the
most commonly used analytical tools for determining (both quali-
tatively and quantitatively) the chemical compositions of a variety
of archaeological and historical materialsdprimarily obsidian (e.g.,
Shackley, 1998) and metals (e.g., Guerra, 1998), but also ceramics
(e.g., Buxeda et al., 2003; Hall, 2004; Iñañez et al., 2007), flint (e.g.,
Hughes et al., 2010), metallurgical slags (e.g., Charlton et al., 2010),
and soils (e.g., Marwick, 2005). Until recently, most archaeological
applications of XRF have been confined to dedicated laboratories
where XRF research has been overseen by scientists knowledgeable
about physics and chemistry. During the past ten years, this model
has changed significantly as archaeologists, museum conservators,
and curators with limited-to-no background in chemistry, physics,
and/or provenance-based studies of archaeological materials, have
begun to acquire and use with increasing frequency commercially
available portable XRF spectrometers. Dubbed portable XRF (PXRF,
pXRF), field-portable XRF (FPXRF), or handheld XRF, such instru-
mentation has been commercially available since the early-to-mid: þ1 301 238 3709.
).
r Ltd.1960s (Piorek, 1997), and was being used to analyze archaeological
materials as early as the 1970s (e.g., Cesareo et al., 1973). By the
mid-1990s, following technological developments in computing
and instrumentation (e.g., tubes, detectors, and associated elec-
tronics), portable XRF instruments were beginning to see increased
use in geology (e.g., Potts et al., 1995, 1997a) and archaeology (e.g.,
Emery and Morgenstein, 2007; Morgenstein and Redmount, 2005;
Pantazis et al., 2002; Potts et al., 1997b; Williams-Thorpe et al.,
1999, 2003). And, since about 2005, there has been a tremendous
increase in the number of portable instruments sold to the art and
archaeological communities. We conservatively estimate there are
approximately 200 such instruments worldwide currently being
used at universities and museums in support of archaeological
research and art conservation science. Despite the commercial
availability of portable XRF analyzers for almost 50 years, the most
common applications have been, and remain today, the metals
recycling and mining industries.
Although portable XRF instruments have proven to be valuable
research tools, the wide-spread adoption of portable XRF is not
without issues. It is abundantly clear to many of us in the field of
archaeological chemistry that a large number of handheld XRF users
lack fundamental knowledge of X-ray spectroscopy, chemistry,
physics, materials science, and/or statistics (see Shackley, 2010a,
2010b for additional discussion). Likewise, many such users have
little or no background in provenance studies of archaeological
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e34963484materials, and fail to grasp the most basic principles that underlie
the provenance postulate (e.g., Neff, 2000; Weigand et al., 1977).
This “black box” approach is inherently problematic to archaeo-
logical science. We underscore that it is not the analytical technique
that is flawed, but rather the general lack of experience and
knowledge of the users of this technology. Certainly this is not the
case for all researchers who use portable XRF, but rather a broad
generalization. We note that that there are no fundamental differ-
ences in the physics or capabilities between portable XRF and more
conventional energy dispersive XRF (EDXRF) instrumentsdboth
instruments types (lab and portable) have X-ray tubes and detectors
that allow elemental information about a given sample to be
generated. The major differences are that portable XRF instruments
are portable; typically lack the ability to measure lower atomic
weight elements under “true” vacuum; objects are typically
measured outside of a chamber using a point and shoot approach.
Additionally, portable instruments generally lack the higher pow-
ered software necessary for the deconvolution of spectra and
quantification of data that is available in most laboratory-based
instruments. Finally lab-based XRF instruments can produce
slightly higher energy X-rays that may allow measurement of an
additional element or two, but these differences are relatively
minor. Hence for quantitative analyses, portable XRF instruments
are ruled by the same constraints as any other laboratory-based XRF
spectrometer. These include among others, clean flat surfaces,
homogeneous samples, and calibration algorithms that ideally are
based on matrix-matched standards and reference materials.
Although many portable XRF instruments come with factory-
installed quantification routines that convert X-ray intensities to
abundance data, these calibrations typically are not comparable to
the types of artifacts being analyzed and oftentimes produce
spurious results. From a qualitative perspective, analysts can use
the XRF spectra to identify the presence and/or absence of various
elements, such as the presence of heavymetals (e.g., As, Hg, and Pb)
used to protect ethnographic objects from insect infestations.
However, quantification by XRF ideally requires that users be
capable of creating their own calibrations and be able to evaluate
these calibrations and resulting data in terms of accuracy, precision,
and sensitivity. Whereas some portable XRF vendors might argue
their preinstalled calibration routines are satisfactory and therefore
this step is not necessary, the fact remains that reliable and accurate
factory calibrations for ceramics and obsidian1 do not exist. We
argue, that in many (if not most) cases, factory-installed calibra-
tions are not adequate for quantification of elements in non-
metallic archaeological materials. Portable XRF users who initiate
projects based on the analyses of ceramics, obsidian, and other
archaeological materials, using untested factory calibrations (i.e.,
calibrations not validated by the user through the analysis of
known reference materials similar in composition to the objects
being analyzed), do so at risk of generating data that are prob-
lematic and/or unusable. With respect to factory calibrations, the
one exception (based on our experience) is that most factory cali-
bration routines for metal alloys tend to produce accurate
resultsdat least for the major elements (e.g., Heginbotham et al.,
2010). However, most archaeological metals are corroded and/or
heavily patinated on the surface. Because XRF involves a shallow
penetration of X-rays into the sample, elements that are enriched
(or depleted) on the surface are predominantly being measured,
which may be significantly different from the bulk composition of
the metal.1 We have recently learned that one portable XRF manufacturer (Bruker) has
obtained a set of ca. 30 well characterized obsidian reference samples for cali-
brating instruments for obsidian analysis.Companies that sell portable XRF spectrometers routinely make
claims that it is possible for individuals to analyzedwith minimal
trainingda wide variety of archaeological, ethnographic, and art
objects, such as ceramics, native copper, paints and pigments, and
man-made glass. These claims are absolutely true! Anyone can be
trained, with little effort, in the mechanics of using such a device to
produce a spectrum. The disconnect lies in that the analysis (in this
case the mechanics of producing a spectrum) is not necessarily the
same as sourcing (in the sense of the term as we have come to
expect). Simply generating numbers and creating compositional
groups is not a sourcing study regardless of the analytical tech-
nique. Likewise, qualitative approaches that involve stacking
spectra on top of one another and pointing out differences in the
spectra is not sourcing either. Researchers who generate (or
incorporate) portable XRF data into their programs should question
how the data compare to data generated by other analytical tech-
niques? Are they reproducible? How do they compare to other
published reference materials and standards? Can the data be used
by other researchers in the future, or are the numbers only
“internally consistent” and therefore only valid for the purposes of
the current study. The same holds true for any analytical tech-
nique. Although we have singled out portable XRF in the above
discussion, these criteria apply to all types of analyses. We under-
score that we are not biased against the use of portable XRF, and in
fact, the senior author of this paper has long been a proponent of
this technology for both qualitative and quantitative analyses of
cultural heritage and geological materials (e.g., Aldenderfer et al.,
2008; Brostoff et al., 2009; Craig et al., 2007, 2010; Cecil et al.,
2007; Farris et al., in press; Glascock et al., 2007a,b; Goebel et al.,
2008; Grissom et al., 2010; Heginbotham et al., 2010; Little et al.,
in press; Phillips and Speakman, 2009; Reuther et al., 2011;
Slobodina et al., 2009; Speakman et al., 2007; Wolff et al., in press).
For decades, laboratory-based XRF analyses of ceramics have
occurred with some degree of regularity, but XRF of ceramics never
has quite achieved the popularity of other analytical methods. The
reason for this, in part, is the reduced sensitivity for trace elements
relative to other analytical techniques such as instrumental neutron
activation analysis (INAA) or inductively coupled plasma-mass
spectrometry (ICP-MS), and generally speaking fewer “potentially
discriminating” elements are analyzed (e.g., rare earth elements). It
is no surprise then that laboratory-based XRF analyses have not
witnessed the same degree of popularity as INAA of ceramics
despite the ubiquitous nature of laboratory-based XRF instruments
at most major universities, whereas only a handful of reactors exist
with research programs dedicated to characterization of archaeo-
logical ceramics. What is surprising is the apparent interest by
archaeologists in using portable XRF to source ceramics (as wit-
nessed by presentations at recent national and international
archaeology and archaeometry meetings). Typically, and under
optimal conditions, a laboratory-based XRF analysis of ceramics can
provide usable data for approximately 20e25 elements versus the
10e20 elements that can reasonably be expected from a portable
XRF analyzer. Traditionally, for quantitative analyses by laboratory-
based XRF, several grams of homogenized ceramic powder typically
is pressed into pellets, fused into disks, or both in order to measure
majors and trace elements separately (e.g., Hein et al., 2002).
The same approach is equally valid for portable XRF instruments,
but does not seem to have been widely adopted by users of
portable instruments. Although portable XRF is theoretically non-
destructive, to correctly analyze ceramics the surface must be
clean of slip, paint, glaze, dirt, etc. At a minimum this means that (1)
the ceramic surface should be abraded with something akin to
a silicon carbide burring tool to prepare the ceramic surface for
analysis or (2) that the samples be prepared as pressed pellets or
fused disks. A cut or broken edge also is a possibility assuming that
Fig. 1. Examples of Mimbres Black-on-white pottery. Photographs courtesy of the Smithsonian National Museum of Natural History.
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e3496 3485the sherd thickness is greater than the diameter of the XRF beam
(ca. 4e10 mm, depending on the instrument) and that the area
being analyzed is relatively flat2 and more or less homogeneous.
Hence, true non-destructive sourcing of whole pots in museum
collections is likely not a reality, although portable XRF may be the
only practical method of acquiring any chemical compositional data
in cases wheremuseums will not permit “destructive” sampling for
INAA or other analytical methods.
In a recent report, Tagle and Gross (2010) described the use of
mXRF for the non-destructive analysis of Mesoamerican and other
ceramics. The specific mXRF used in Tagle’s study had a spot size of
ca. 3 mm diameter, which is not considerably different from the
3e5 mm diameter spot size of some commercial portable XRF
instrumentsdincluding the instrument used herein. However, the
idea of using a “true” mXRF spectrometer (e.g., an instrument with
a beam size of ca. 30e100 mm diameter, or smaller) to source
ceramics is troubling. At the 30e100 mm level of analysis what
exactly is being measureddthe clay, non-plastics, or a combination
of both? As part of this research, we conducted a limited mXRF study
(see below) to illustrate the effects of homogeneity and heteroge-
neity at the mm level.
Despite the challenges and limitations of using portable XRF to
source ceramics, the possibility for sourcing pottery remains,2 With calibration schemes that involve ratioing counts to the Compton peak or
Bremsstrahlung continuum sample “flatness” is less of a problem in EDXRF. Flatness
is a major concern when fundamental parameters (FP) calibrations are being used
given that the calculations are based on the theoretical relationships between X-ray
intensities and element concentrations in an ideal sample.provided that there is adequate sample preparation, i.e., removal of
soil, slip, pigments, etc. from the sherd. In a recent paper, Goren and
colleagues (Goren et al., 2011) concluded that portable XRF
instruments could be used to nondestructively source clay tablets
in cases where internal groupings had previously been established
(e.g., by INAA). The authors correctly stressed that portable XRF was
not a substitute for other chemical analyses or petrography, but that
portable XRF could be useful for sourcing clay tablets when intru-
sive sampling is not possible. In this paper, we explore the possi-
bility of using portable XRF to source Mimbres and Jornada pottery
from the American Southwest. Unlike the Goren et al. study which
was conducted on what can be considered fine paste ceramics, we
focused on pottery containing varying amounts of temper.
2. Samples
2.1. mXRF samples
mXRF analyses were conducted on three samples: (1) an
archaeological pottery specimen, MRM049, which has a moderate
amount of coarsely crushed igneous rock; (2) a fine paste ceramic,
New Ohio Red Clay; and (3) a sample of obsidian from Wiki Peak,
Alaskada theoretically homogeneous material.
2.2. Portable XRF samples
Since publication of Gilman et al.’s (1994) seminal INAA study of
Mimbres pottery (Fig. 1), dozens of research projects involving the
analyses of Mimbres and Jornada pottery have been conducted at
Table 1
Descriptive information for pottery samples analyzed by portable XRF and INAA.
Anid Group Ceramic Type Site Name Site #
ANI022 Mimbres-21 Mimbres B/w Style III AZ BB1635
ANI024 Mimbres-21 Mimbres B/w Style III AZ U15127
ANI026 Mimbres-21 Mimbres B/w Style III AZ U15127
ANI029 Mimbres-21 Mimbres B/w Style III AZ BB1641
ANI030 Mimbres-21 Mimbres B/w Style III AZ BB1641
MRM003 Mimbres-11 Mimbres B/w Style III Hueco Tanks 41EP00002
MRM006 Mimbres-11 Mimbres B/w Style II Hueco Tanks 41EP00002
MRM026 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM028 Mimbres-04a Mimbres B/w Style III North Hills I 41EP00355
MRM033 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM035 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM042 Mimbres-01 Mimbres B/w Style III North Hills I 41EP00355
MRM043 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM049 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM050 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM052 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM053 Mimbres-02a Mimbres B/w North Hills I 41EP00355
MRM054 Mimbres-02a Mimbres B/w Style III North Hills I 41EP00355
MRM057 Mimbres-04a Mimbres B/w Style III North Hills I 41EP00355
MRM062 Mimbres-11 Mimbres B/w Style III North Hills I 41EP00355
MRM063 Mimbres-01 Mimbres B/w Style III North Hills I 41EP00355
MRM077 Mimbres-01 Mimbres B/w Style III Divad LA 096687
MRM081 Mimbres-11 Mimbres B/w Style III Divad LA 096687
MRM086 Mimbres-01 Mimbres B/w Style III Divad LA 096687
MRM087 Mimbres-02a Mimbres B/w Style III Divad LA 096687
MRM090 Mimbres-11 Mimbres B/w Style III Divad LA 096687
MRM094 Mimbres-11 Mimbres B/w Style III Divad LA 096687
MRM104 Mimbres-04a Mimbres B/w Style II Turq Ridge FB6307
MRM115 Mimbres-01 Mimbres B/w Style III Sandcliffe Sandcliffe
MRM116 Mimbres-11 Mimbres B/w Style III Sandcliffe Sandcliffe
MRM126 Mimbres-01 Mimbres B/w Style III Temporal LA 001085
MRM128 Mimbres-11 Mimbres B/w Style III Temporal LA 001085
MRM130 Mimbres-01 Mimbres B/w Style III Temporal LA 001085
MRM131 Mimbres-01 Mimbres B/w Style III Temporal LA 001085
MRM146 Mimbres-04a Mimbres B/w Style II Roth LA 073942
MRM150 Mimbres-03 Mimbres B/w Style II Roth LA 073942
MRM166 Mimbres-11 Mimbres B/w Style III Los Tules LA 016315
MRM191 Mimbres-08 Mimbres B/w Style II Los Tules LA 016315
MRM276 El Paso Core El Paso Polychrome Hueco Tanks 41EP00002
MRM282 El Paso Core El Paso Bichrome North Hills I 41EP00355
MRM286 El Paso Core El Paso Brownware Diablo 1 41HZ0491
MRM289 El Paso Core El Paso Brownware Diablo 3 41HZ0493
MRM290 El Paso Core El Paso Brownware Conejo LA 091044
MRM295 El Paso Core El Paso Brownware Hill 100 LA 097088
MRM305 El Paso Core El Paso Brownware Roth LA 073942
MRM307 El Paso Core El Paso Brownware Roth LA 073942
MRM311 El Paso Core El Paso Brownware Los Tules LA 016315
MRM312 El Paso Core El Paso Brownware Los Tules LA 016315
OT102 Mimbres-08 Mimbres B/w Style III LA 018342
OT157 Mimbres-08 Mimbres B/w Style III Ronnie Pueblo 83-NM-400
OT174 Mimbres-11 Mimbres B/w, Style II Old Town LA 001113
OT187 Mimbres-21 Mimbres B/w, Style II Old Town LA 001113
OT189 Mimbres-21 Mimbres B/w, Style II Old Town LA 001113
OT195 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT196 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT197 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT198 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT199 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT200 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT205 Mimbres-21 Mimbres B/w, Style II Old Town LA 001113
OT230 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT231 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT233 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT234 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT235 Mimbres-21 Mimbres B/w, Style II Old Town LA 001113
OT243 Mimbres-03 Untyped, Unslipped B/w Old Town LA 001113
OT504 Mimbres-08 Mimbres B/w Style III Avilas Canyon Village LA 045000
OT506 Mimbres-08 Mimbres B/w Style III Avilas Canyon Village LA 045000
OT511 Mimbres-04a Mimbres B/w Style III Avilas Canyon Village LA 045000
OT512 Mimbres-04a Mimbres B/w Style III Avilas Canyon Village LA 045000
OT513 Mimbres-04a Mimbres B/w Style III Avilas Canyon Village LA 045000
WCRM003 Mimbres-04a Mimbres B/w Style III Badger Ruin LA 111395
WCRM004 Mimbres-04a Mimbres B/w Style III Badger Ruin LA 111395
WCRM007 Mimbres-11 Mimbres B/w Style III Jackson Fraction LA 111413
WCRM010 Mimbres-08 Mimbres B/w Style III Jackson Fraction LA 111413
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e34963486
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e3496 3487Texas A&M and the University of Missouri Research Reactor Center
(MURR). These projects have ranged in scale from small cultural
resource management projects involving the analyses of relatively
few samples to large-scale research projects initiated by Mimbres
scholars. Most of the early work involving Mimbres pottery was
conducted at Texas A&M in the early-to-mid 1990s by nuclear
chemist Dennis James in collaboration with archaeologist Harry
Shafer, Robbie Brewington, Holly Meier, and Eleanor Dahlin. Under
the Texas A&M program, more than 1100 Mimbres, Jornada, and
related pottery and clays were analyzed. Since the mid-1990’s,
archaeologists Myles Miller, Darrell Creel, and numerous other
researchers (e.g., Pat Gilman, Nancy Kenmotsu, Lori Reed, Bernard
Schriever, Chris Turnbow, and others) have initiated dozens of
INAA-based research projects through the MURR laboratory that
were complementary to the Texas A&M program and Gilman’s
earlier research through the Smithsonian-NIST INAA program. As
a result of these combined efforts, more than 4000 Mimbres, Jor-
nada, and related pottery specimens and clay samples from more
than 250 sites throughout southern New Mexico, eastern Arizona,
west Texas, and Chihuahua have been analyzed to date. Although
these subsequent studies confirmed the patterns reported by
Gilman et al. (1994) and have produced evidence for local ceramic
production at a number of sites in the Mimbres Valley and else-
where, there has been no published synthesis of the data. An
ongoing major analytical effort, by Speakman, Creel, and Miller,
focused on all extant INAA data has resulted in the identification of
at least 45 Mimbres, Jornada, and related compositional groups.
Interpretation and publication of these data is ongoing.Fig. 2. Map of eastern Arizona, southwest New Mexico, and west Texas showing the site lo
sitional groups discussed in the text.For this study we selected 75 Mimbres and Jornada pottery
samples representing eight of what we believe to be the most
clearly defined and chemically distinct of the 45 compositional
groups. Each sample selected for portable XRF analysis had previ-
ously been analyzed by INAA at MURR allowing us to directly
compare the results.
The eight compositional groups represented in the portable XRF
sample include Mimbres-1, Mimbres-2a, Mimbres-3, Mimbres-4a,
Mimbres-8, Mimbres-11, Mimbres-21, and El Paso Core. Descriptive
information and group assignments (based on INAA) for each
sample are presented in Table 1. With the exception of the samples
assigned to the El Paso Core group which represents Jornada
brownwares produced in the vicinity of El Paso, Texas, most pottery
is typologically classified as Mimbres Black-on-white Style III and
dates to ca. AD 1000e1150.
Despite the large numbers of Mimbres pottery analyzed by
INAA, the assignment of the compositional groups to a specific site
or groups of sites is problematic for several reasons. The first is
that the criterion of abundance (Bishop et al., 1982) simply does not
apply to Mimbres pottery. In most cases Mimbres pottery
(including plain wares) was so widely moved around the landscape
that at best only 20e30% of pottery analyzed from a given site
might be considered locally produced. Consequently, we cannot
confidently attribute compositional groups to a specific site or sites
based on numbers alone. In addition, relatively few clays (<150)
have been analyzed, and fewer than five compositional groups can
be confidently linked to specific sites based on clay chemistry.
Finally, the larger dataset of pottery represents numerous projects,cations of pottery analyzed in this study and the productions areas for the 8 compo-
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e34963488each with different research questions. As a result, the sample is
highly biased with approximately 15 sites representing ca. 50% of
the sample. Having said this, we underscore that the point of this
specific experiment is not to source the ceramics (we’ve already
done that). Instead, our primary purpose here is to evaluate the
potential of portable XRF for constructing compositional groups
that can be used for provenance-based studies of ceramics.Fig. 3. mXRF line scans for selected elements showing chemical variability in Ohio Red Clay
Black Line). A total of 130 individual measurements are represented by each line. Measureme
a helium atmosphere. Note: Cr and Mg are below detection limits in the obsidian sample. (Fo
the web version of this article.)The eight compositional groups selected for this study represent
pottery production in five discrete geographic areas (Fig. 2). As
indicated above, pottery assigned to the El Paso Core group repre-
sents pottery production in El Paso, Texas. The Mimbres-1 group
represents pottery production in the Rio Grande Valley of southern
New Mexico, probably the Rio Vista site and possibly the area
immediately north of Rio Vista. Mimbres-3 pottery was(Red line), archaeological pottery (MRM049; Blue line), and obsidian (Wiki Peak, AK;
nts occurred at 80 mm intervals, at 50 kV and 600 mA for a live-time count of 60 s each in
r interpretation of the references to colour in this figure legend, the reader is referred to
Fig. 3. (continued).
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e3496 3489manufactured in the Gila Forks area of New Mexico. Unlike other
Mimbres Black-on-white pottery which is recognizable as a white-
slipped brownware, pottery assigned to Mimbres-3 has a white
paste. Mimbres-21 pottery was produced at the Woodrow site
which is located along the Gila River in New Mexico, near the Ari-
zona border. The remaining four groups, Mimbres-2a, -4a, -8, and
-11 originate from the Mimbres Valley in southwest New Mexico.
Mimbres-2a pottery was likely produced in the middle part of the
Mimbres Valley, probably at Swarts Ruin. Mimbres-4a pottery is
probably from the Galaz site, and Mimbres-8 is believed to repre-
sent pottery production at the Mattocks site. Finally, Mimbres-11currently is believed to represent pottery production at Pruitt
Ranch.
3. Analytical methods
3.1. mXRF sample preparation and analysis
The archaeological pottery sample was abraded using a silicon
carbide burr to remove the slip, paint, and/or any other materials
adhering to the surface. The samplewas washed in deionized water
and allowed to dry. The Ohio Red Clay sample is a test tile that was
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e34963490produced at MURR and fired to 1100 C (Cogswell et al., 1996).
Because the sample was flat and had been stored in a clean plastic
bag since its manufacture, no sample preparation was necessary.
The Wiki Peak obsidian sample was prepared by using a lapidary
sawwith a diamond blade to remove a ca. 2 2 0.5 cm slice from
a larger geological specimen. The sample subsequently was washed
with deionized water.
mXRF analyses were conducted using a Bruker AXS ARTAX 800
spectrometer equipped with a rhodium target polycapillary lens X-
ray tube that has ca. 80 mm spatial resolution. The X-ray detector is
a Si drift detector (SDD) with a 10 mm2 active area and energy
resolution of ca. 142 eV for the Mn Ka peak at 100 kcps. For this
experiment, the instrument was programmed to travel along a line
ca. 10.4 mm in length. A total of 130 individual measurements
occurred on each sample, at 80 mm intervals, at 50 kV and 600 mA,
for a live-time count of 60 s at each spot. Analyses were conducted
in a helium atmosphere to maximize Mg, Al, and Si measurements.3.2. INAA sample preparation and irradiation
INAA sample preparation and analysis were conducted at the
University of Missouri Research Reactor Center (MURR). The
pottery samples were prepared for INAA using standard MURR
procedures (Glascock et al., 2007a,b). Fragments of about 1e2 cm
were removed from each sample and the surfaces were abraded
using a silicon carbide burr to remove the slip, paint, and/or any
other materials adhering to the surface, thereby reducing the risk of
measuring surface contamination. The samples were washed in
deionized water and allowed to dry in the laboratory. Once dry, the
individual sherdswere ground to a fine homogenized powder using
an agate mortar and pestle.
Aliquots of approximately 150 mg of powder were weighed into
small polyvials used for short irradiations at MURR. At the same
time, 200 mg of each sample were weighed into the high-purity
quartz vials used for long irradiations. Along with the unknown
samples, reference standards of SRM-1633a (coal fly ash) and SRM-
688 (basalt rock) were similarly prepared, as were quality control
samples (e.g., standards treated as unknowns) of SRM-278
(obsidian rock) and Ohio Red Clay.Table 2
mXRF data summary statistics for MRM049, Ohio Red clay, and Wiki Peak obsidian. %rsd
MRM049 Ohio Red Clay (N
Average Counts std %rsd Average Counts
Mg 145 93 64 361
Al 25,642 3980 16 25,513
Si 174,698 25,104 14 144,248
K 89,190 28,381 32 101,414
Ca 33,225 16067 48 8494
Ti 37,917 16,162 43 63,744
Cr 1111 1246 112 2227
Mn 4512 1961 43 6765
Fe 817,290 199,095 24 1,679,301
Ni 1142 2484 217 2900
Cu 1659 1331 80 1063
Zn 6479 1895 29 6529
Ga 2340 1702 73 833
Rb 16,868 2555 15 13,383
Sr 13,138 4282 33 4658
Y 207 548 265 506
Zr 15,525 7848 51 14,443
Nb 1197 436 36 880
std e standard deviation.
%rsd e % relative standard deviation.
bdl e below detection limits.INAA of ceramics at MURR, which consists of two irradiations
and a total of three gamma counts, constitutes a superset of the
procedures used at most other NAA laboratories (Glascock, 1992;
Neff, 2000). As discussed in detail by Glascock (1992), a short
irradiation entails the irradiation of each sample for 5 s by a neutron
flux of 8  1013 cm2 s1. The 720-second count yields gamma
spectra containing peaks for short-lived elements: Al, Ba, Ca, Dy, K,
Mn, Na, Ti, and V. The samples encapsulated in quartz vials are
subjected to a 24-h irradiation at a neutron flux of 51013 cm2 s1.
This long irradiation is analogous to the single irradiation utilized at
most other laboratories. After the long irradiation, samples decay
for seven days, and then are counted for 2000 s (the “middle
count”) on a high-resolution germanium detector coupled to an
automatic sample changer. The middle count yields determinations
of seven medium half-life elements, namely As, La, Lu, Nd, Sm, U,
and Yb. After an additional three- to four-week day, a final count of
9000 s is carried out on each sample. The latter measurement yields
the following 17 long half-life elements: Ce, Co, Cr, Cs, Eu, Fe, Hf, Ni,
Rb, Sb, Sc, Sr, Ta, Tb, Th, Zn, and Zr.
3.3. Portable XRF sample preparation and analysis
Portable XRF analyses were conducted on archived sherds
retained from the earlier INAA studies. Ceramic fragments of
approximately 1e2 cmwere abraded using a silicon carbide burr to
flatten the surfaces and remove slip, paint, and/or any other mate-
rials adhering to the surface. The samples were washed in deionized
water and allowed to dry.
Analyses were conducted using a Bruker Tracer III-V handheld
XRF spectrometer equipped with a rhodium tube and a Si-PIN
detector with a resolution of ca. 170 eV FWHM for 5.9 keV X-rays
(at 1000 counts per second) in an area 7 mm2. The spot size on this
specific instrument is ca. 4 mm diameter.
Each sample was measured twice using different excitation
conditions. For the first analysis, each sample was analyzed at
40 kV, 15 mA, with a 0.076-mm copper filter and 0.0305-mm
aluminum filter placed in the X-ray path for a 200-second live-
time count. Peak intensities for the Ka peaks of Mn, Fe, Rb, Sr, Y,
Zr, Nb, and La peak of Th were calculated as ratios to the Compton
peak of rhodium, and converted to parts-per-million (ppm) using(relative standard deviation); Conc. (concentration).
OR) Wiki Peak Obsidian
std %rsd Average Counts std %rsd
101 28 bdl
1776 7 18,819 633 3
8320 6 211,946 5538 3
6439 6 95,718 878 1
2118 25 25,141 2013 8
17,377 27 14,316 196 1
1421 64 bdl
2259 33 9052 208 2
82,389 5 300,527 8617 3
814 28 37 65 174
395 37 866 248 29
841 13 2508 232 9
300 36 2247 190 8
576 4 10,073 270 3
384 8 7960 309 4
1208 238 38 146 380
8051 56 12,397 328 3
296 34 554 228 41
Table 3
INAA and portable XRF data for individual specimens.
Group Al
(PXRF) %
Al
(NAA) %
K
(PXRF) %
K
(NAA) %
Ca
(PXRF) %
Ca
(NAA) %
Ti
(PXRF) %
Ti
(NAA) %
Mn
(PXRF)
ppm
Mn
(NAA)
ppm
Fe
(PXRF) %
Fe
(NAA) %
Th
(PXRF)
ppm
Th
(NAA)
ppm
Rb
(PXRF)
ppm
Rb
(NAA)
ppm
Sr
(PXRF)
ppm
Sr
(NAA)
ppm
ZR
(PXRF)
ppm
Zr
(NAA)
ppm
Y
(PXRF)
ppm
Nb
(PXRF)
ppm
El Paso Core
MRM276 9.07 8.76 3.45 4.40 1.23 1.46 0.53 0.67 563 555 4.44 3.72 14 17 98 99 379 305 399 506 39 38
MRM282 8.66 8.71 2.26 3.11 0.86 1.27 0.48 0.55 517 414 4.41 3.85 13 19 103 113 198 322 421 367 28 28
MRM286 8.73 9.27 2.36 3.50 0.82 1.15 0.36 0.36 440 303 2.84 3.05 16 28 112 138 361 322 308 309 24 31
MRM289 9.32 9.13 1.78 2.94 2.34 1.51 0.46 0.65 627 536 4.73 4.75 20 24 99 100 541 432 483 287 28 41
MRM290 9.24 9.11 2.07 3.10 1.00 1.36 0.40 0.56 453 485 3.80 4.30 12 21 116 112 343 455 361 296 28 31
MRM295 8.45 9.09 2.35 3.21 1.08 1.31 0.38 0.45 847 454 3.40 3.63 12 19 132 134 340 360 303 275 31 26
MRM305 9.59 9.24 3.20 4.05 0.97 1.33 0.56 0.60 625 564 3.56 4.24 14 19 109 112 538 414 544 418 35 38
MRM307 10.11 8.60 2.66 3.28 1.24 1.59 0.38 0.43 604 553 4.33 4.05 12 18 128 109 263 455 338 283 32 22
MRM311 9.78 9.34 2.91 2.82 0.95 1.39 0.35 0.34 663 372 4.63 4.12 21 20 137 132 407 408 298 282 35 33
MRM312 9.37 8.18 2.61 2.91 1.05 1.20 0.42 0.49 610 396 3.90 4.02 13 21 102 105 397 450 295 220 22 31
Mean 9.23 8.94 2.57 3.33 1.15 1.36 0.43 0.51 595 463 4.01 3.97 15 21 114 115 377 392 375 324 30 32
Standard Dev. 0.52 0.37 0.51 0.52 0.44 0.14 0.07 0.11 116 90 0.61 0.45 3 3 14 14 106 60 86 84 5 6
%rsd 5.6 4.1 20.0 15.5 38.0 10.2 16.5 22.2 19.5 19.5 15.2 11.4 21.3 15.2 12.6 12.2 28.2 15.2 23.0 25.8 17.0 18.2
Mimbres-01
MRM042 10.18 9.02 3.01 3.73 0.47 0.58 0.26 0.22 341 418 1.62 1.74 34 36 239 259 136 145 157 149 33 18
MRM063 10.14 9.69 2.49 3.30 0.43 0.63 0.20 0.18 135 127 1.19 1.27 36 46 184 205 79 73 145 193 27 18
MRM077 10.16 10.48 2.90 3.47 0.49 0.65 0.12 0.11 189 165 0.94 1.05 48 56 214 232 82 88 173 166 28 22
MRM086 12.11 9.83 2.78 2.72 1.04 0.78 0.16 0.11 333 149 1.08 1.11 40 50 186 195 115 110 159 177 27 23
MRM115 12.54 10.69 2.85 3.01 0.83 0.76 0.18 0.15 290 305 1.41 1.37 62 55 231 241 115 97 140 155 44 25
MRM126 12.00 9.43 2.92 3.43 1.22 0.75 0.39 0.11 156 170 0.82 0.97 39 49 178 196 152 203 113 170 24 22
MRM130 10.90 9.48 2.92 2.89 0.72 0.59 0.45 0.16 205 153 1.02 1.24 37 49 196 221 140 135 158 166 25 22
MRM131 9.68 9.39 2.74 3.42 0.39 0.78 0.19 0.16 172 171 0.82 1.02 40 47 188 195 130 109 132 147 28 24
Mean 10.96 9.75 2.83 3.25 0.70 0.69 0.24 0.15 228 207 1.11 1.22 42 48 202 218 119 120 147 165 30 22
Standard Dev. 1.10 0.57 0.16 0.34 0.31 0.09 0.12 0.04 82 101 0.28 0.25 9 6 23 24 27 41 19 15 7 2
%rsd 10.0 5.8 5.6 10.5 44.2 12.7 47.5 27.5 35.9 48.7 25.6 20.5 21.4 12.9 11.5 11.1 22.4 34.2 12.8 9.3 22.1 11.1
Mimbres-02a
MRM026 9.58 9.01 2.45 2.79 0.96 1.24 0.30 0.29 479 395 2.63 2.64 21 21 146 159 301 403 227 151 32 20
MRM033 10.18 9.17 2.38 2.83 0.75 1.16 0.32 0.36 473 315 2.69 2.75 16 20 136 155 277 303 319 194 32 21
MRM035 10.43 9.24 2.20 2.55 0.82 1.27 0.35 0.34 343 315 2.78 2.72 18 20 149 155 284 333 232 163 33 18
MRM043 9.02 8.89 2.25 2.68 0.85 1.15 0.40 0.33 378 251 2.60 2.59 17 19 156 147 299 294 216 154 39 21
MRM049 9.61 9.19 2.54 3.29 0.45 0.61 0.34 0.30 200 193 2.08 2.04 20 24 183 190 149 153 262 173 32 23
MRM050 10.21 9.08 2.40 2.86 0.82 1.17 0.32 0.33 419 364 2.35 2.38 14 19 142 151 367 382 250 180 32 20
MRM052 9.91 9.20 2.16 2.60 0.77 1.22 0.32 0.33 434 299 2.70 2.69 21 20 145 144 283 321 268 218 32 18
MRM053 12.61 8.59 1.89 2.68 3.34 1.26 0.29 0.27 396 320 2.49 2.74 16 19 140 143 279 375 208 169 30 17
MRM054 10.00 9.67 2.19 2.71 0.70 0.98 0.31 0.33 439 356 2.48 2.59 18 20 150 155 328 348 232 181 34 20
MRM087 10.07 8.40 2.31 2.76 0.86 1.21 0.36 0.31 513 533 2.58 2.62 15 20 144 152 335 403 253 187 36 21
Mean 10.16 9.04 2.28 2.77 1.03 1.13 0.33 0.32 407 334 2.54 2.58 17 20 149 155 290 331 247 177 33 20
Standard Dev. 0.95 0.35 0.18 0.21 0.82 0.20 0.03 0.03 88 90 0.21 0.22 2 1 13 13 58 74 32 20 2 2
%rsd 9.3 3.9 8.0 7.4 79.6 17.7 9.7 8.2 21.7 27.1 8.1 8.3 13.4 6.9 8.8 8.7 19.8 22.3 13.0 11.3 7.5 9.0
Mimbres-03
MRM150 9.03 8.14 2.51 4.70 0.98 0.82 0.27 0.11 690 535 1.36 1.38 41 38 281 289 100 90 240 193 167 41
OT195 9.74 7.98 5.43 5.24 0.49 0.88 0.12 0.11 655 713 1.16 1.23 42 37 289 304 40 42 209 226 76 29
OT196 11.31 8.39 4.91 5.02 0.71 1.18 0.11 0.15 401 667 0.98 1.35 39 36 277 283 60 63 215 201 66 29
OT197 10.52 8.48 5.12 5.14 0.73 0.29 0.11 0.07 504 672 0.96 1.37 47 37 279 291 69 54 207 206 85 31
OT198 10.07 8.03 4.91 5.05 0.53 0.81 0.11 0.08 557 538 1.06 1.31 39 32 271 278 68 67 183 167 66 27
OT199 10.85 8.31 5.13 5.31 0.53 0.68 0.13 0.06 420 614 1.03 1.39 31 36 300 305 59 49 223 194 113 32
OT200 11.16 8.14 5.31 5.58 0.99 1.26 0.12 0.19 508 651 1.22 1.47 44 35 315 329 52 50 187 186 81 27
OT230 10.30 8.32 4.35 4.56 0.74 1.08 0.13 0.13 696 622 1.22 1.26 38 35 250 258 90 90 275 156 80 30
OT231 11.25 7.84 5.12 5.41 0.92 1.65 0.15 0.23 572 630 1.04 1.22 36 34 280 282 63 68 197 151 103 29
OT233 12.37 8.24 5.95 5.19 0.90 0.87 0.15 0.14 556 663 1.02 1.29 49 37 322 312 74 58 212 187 106 34
OT234 11.25 8.66 5.45 5.01 0.84 1.11 0.11 0.14 469 613 0.98 1.21 44 36 271 286 49 39 194 162 78 31
(continued on next page)
R.J.Speakm
an
et
al./
Journal
of
A
rchaeological
Science
38
(2011)
3483
e
3496
3491
Table 3 (continued )
Group Al
(PXRF) %
Al
(NAA) %
K
(PXRF) %
K
(NAA) %
Ca
(PXRF) %
Ca
(NAA) %
Ti
(PXRF) %
Ti
(NAA) %
Mn
(PXRF)
ppm
Mn
(NAA)
ppm
Fe
(PXRF) %
Fe
(NAA) %
Th
(PXRF)
ppm
Th
(NAA)
ppm
Rb
(PXRF)
ppm
Rb
(NAA)
ppm
Sr
(PXRF)
ppm
Sr
(NAA)
ppm
ZR
(PXRF)
ppm
Zr
(NAA)
ppm
Y
(PXRF)
ppm
Nb
(PXRF)
ppm
OT243 12.16 8.37 5.39 4.71 0.70 0.64 0.14 0.18 471 604 1.08 1.33 33 35 263 269 89 54 218 174 94 33
Mean 10.83 8.24 4.97 5.08 0.75 0.94 0.14 0.13 542 627 1.09 1.32 40 36 283 290 68 60 213 183 93 31
Standard Dev. 0.96 0.23 0.86 0.30 0.18 0.35 0.04 0.05 99 52 0.12 0.08 5 1 21 19 18 16 25 22 28 4
%rsd 8.9 2.8 17.4 6.0 23.4 36.9 32.1 38.6 18.3 8.4 11.2 6.0 13.4 3.9 7.3 6.7 26.5 27.3 11.7 12.2 29.9 11.9
Mimbres-04a
MRM028 9.13 9.11 1.98 2.69 0.95 1.35 0.36 0.25 516 493 3.06 2.70 18 17 128 137 228 301 296 189 30 19
MRM057 9.60 8.83 1.95 2.79 1.01 1.35 0.30 0.26 540 519 2.86 2.83 17 22 124 126 247 246 237 199 31 20
MRM146 9.47 7.86 2.69 3.20 0.89 1.32 0.14 0.22 570 419 2.46 2.69 20 21 181 181 234 241 225 198 29 16
OT511 8.97 8.80 2.32 2.70 0.61 1.00 0.39 0.36 292 379 2.36 2.47 15 19 134 150 222 257 366 266 27 17
OT512 8.55 8.82 2.31 2.96 0.66 1.10 0.29 0.29 282 203 1.64 1.98 16 20 135 147 246 288 191 180 30 15
OT513 8.31 7.74 2.36 2.78 0.59 0.97 0.35 0.39 329 308 2.30 2.29 16 19 143 153 238 251 271 274 29 17
WCRM003 9.67 8.35 2.35 2.82 0.76 1.01 0.34 0.29 203 248 2.17 2.28 18 19 133 140 251 249 236 168 28 15
WCRM004 8.74 8.99 2.28 2.93 0.53 1.30 0.29 0.27 723 376 2.36 2.37 19 27 174 177 136 234 200 169 38 18
MRM104 9.35 8.97 1.93 2.62 0.83 1.37 0.39 0.39 443 390 3.15 3.08 10 18 113 120 275 303 246 184 28 18
Mean 9.09 8.61 2.24 2.83 0.76 1.20 0.32 0.30 433 371 2.49 2.52 17 20 141 148 231 263 252 203 30 17
Standard Dev. 0.48 0.50 0.25 0.18 0.17 0.17 0.08 0.06 169 104 0.47 0.34 3 3 23 21 39 26 54 40 3 2
%rsd 5.3 5.8 11.1 6.3 22.5 14.3 24.5 21.1 38.9 28.1 19.0 13.4 17.7 15.0 16.1 13.9 16.8 10.1 21.3 19.5 10.8 10.2
Mimbres-08
MRM191 8.34 8.46 2.62 3.31 0.56 0.77 0.30 0.30 323 183 1.78 2.11 22 27 193 198 158 152 231 173 27 19
OT102 8.45 8.39 2.38 2.75 0.38 0.61 0.27 0.24 288 195 1.86 2.16 21 31 209 199 114 131 215 165 31 16
OT157 7.71 8.30 2.42 2.92 0.48 0.85 0.26 0.28 399 248 1.99 2.10 22 28 191 193 140 584 201 197 28 16
OT504 8.71 8.49 2.35 2.90 0.45 0.62 0.30 0.28 242 201 1.90 2.15 22 27 156 189 120 119 186 198 25 15
OT506 9.08 8.17 2.83 3.29 0.39 0.60 0.30 0.32 416 497 1.87 1.89 21 26 199 196 130 150 274 208 31 19
WCRM010 9.14 8.07 2.66 3.14 0.54 0.74 0.27 0.22 479 330 1.83 2.03 25 27 198 199 121 112 224 175 27 17
Mean 8.57 8.32 2.54 3.05 0.46 0.70 0.28 0.27 358 275 1.87 2.08 22 28 191 196 131 208 222 186 28 17
Standard Dev. 0.53 0.17 0.19 0.23 0.07 0.10 0.02 0.04 88 121 0.07 0.10 2 2 18 4 16 185 30 17 3 2
%rsd 6.2 2.0 7.5 7.5 15.6 15.0 7.2 13.4 24.7 43.9 3.8 4.8 7.0 5.8 9.5 2.0 12.6 89.0 13.6 9.3 9.2 10.0
Mimbres-11
MRM003 10.87 7.85 2.85 2.97 0.88 0.92 0.37 0.26 769 996 2.35 2.15 20 18 191 198 241 199 204 157 32 21
MRM006 10.80 8.48 2.64 3.13 1.12 1.28 0.36 0.38 482 397 2.06 2.08 16 20 177 184 236 221 225 172 32 22
MRM062 10.85 8.90 2.81 3.26 0.76 1.07 0.40 0.36 1134 1088 2.47 2.35 17 19 189 189 234 265 213 143 32 19
MRM081 10.26 8.26 2.66 2.92 1.07 0.71 0.30 0.20 3068 2331 2.11 2.19 18 20 190 200 225 243 191 121 34 18
MRM090 11.76 8.80 2.63 2.89 0.92 0.85 0.37 0.23 929 1155 3.13 3.02 14 19 190 190 190 177 200 165 29 17
MRM094 13.34 8.20 2.81 3.23 2.36 1.09 0.36 0.42 780 741 2.28 2.50 15 19 204 198 224 260 215 165 31 22
MRM116 12.54 8.39 2.85 2.89 0.83 1.05 0.18 0.30 1229 1384 2.52 2.64 15 19 191 199 246 294 212 161 29 19
MRM128 10.41 8.86 2.92 4.15 0.54 1.06 0.12 0.33 2109 2367 2.34 2.41 17 21 202 208 232 223 191 159 32 21
MRM166 11.05 8.13 2.90 3.27 1.02 0.91 0.38 0.37 965 796 1.95 2.21 13 20 181 184 271 247 254 169 28 23
OT174 11.33 8.75 2.90 2.52 0.79 1.00 0.39 0.36 401 338 2.09 2.25 28 22 197 191 233 243 235 155 34 22
WCRM007 11.87 9.02 2.49 3.11 0.86 1.07 0.42 0.59 1405 1663 2.81 2.77 17 18 181 174 262 253 209 136 29 22
Mean 11.37 8.51 2.77 3.12 1.01 1.00 0.33 0.35 1206 1205 2.37 2.42 17 20 190 192 236 239 214 155 31 21
Standard Dev. 0.94 0.38 0.14 0.40 0.47 0.15 0.10 0.11 775 686 0.35 0.29 4 1 9 10 21 32 19 16 2 2
%rsd 8.2 4.4 5.1 12.9 46.8 15.1 28.6 30.6 64.2 56.9 14.7 12.2 23.3 6.4 4.5 5.0 9.0 13.6 8.8 10.1 6.8 9.0
Mimbres-21
ANI022 10.48 9.85 2.80 3.07 0.72 0.99 0.33 0.30 851 927 2.63 2.78 9 11 144 153 275 317 246 158 30 11
ANI024 10.81 8.61 2.59 3.02 0.91 1.16 0.29 0.25 771 885 2.38 2.45 8 10 129 133 327 356 258 153 26 14
ANI026 11.27 9.36 2.46 2.34 1.08 1.22 0.31 0.25 904 1180 3.12 3.28 8 9 117 119 374 444 232 147 24 11
ANI029 11.90 9.82 2.61 2.87 1.03 1.36 0.28 0.27 955 1076 2.51 2.91 9 10 121 130 372 385 241 160 25 13
ANI030 10.39 8.78 2.23 2.37 0.81 1.23 0.25 0.25 1105 1225 2.27 2.63 7 10 114 117 368 411 217 148 23 11
OT187 12.49 8.63 2.80 2.75 1.71 1.79 0.23 0.28 734 1073 1.97 2.44 10 10 98 110 351 403 187 147 22 8
OT189 11.83 9.15 2.53 2.57 1.62 2.37 0.23 0.15 912 1145 2.28 2.47 10 9 107 114 400 480 275 143 23 12
OT205 10.37 9.08 2.36 2.72 0.93 1.16 0.28 0.27 988 1154 2.88 3.13 11 10 133 131 354 362 229 172 24 13
OT235 11.34 8.93 2.56 2.84 0.84 1.46 0.27 0.18 1034 1028 2.22 2.39 14 9 126 128 380 265 233 140 24 13
Mean 11.21 9.13 2.55 2.73 1.07 1.41 0.27 0.24 917 1077 2.47 2.72 10 10 121 126 356 380 235 152 25 12
Standard Dev. 0.76 0.47 0.19 0.26 0.35 0.42 0.04 0.05 120 114 0.36 0.33 2 1 14 13 36 65 25 10 2 2
%rsd 6.8 5.1 7.3 9.5 32.9 30.0 12.9 19.7 13.1 10.6 14.4 12.0 20.5 7.2 11.4 10.2 10.2 17.1 10.6 6.7 8.9 13.1
R.J.Speakm
an
et
al./
Journal
of
A
rchaeological
Science
38
(2011)
3483
e
3496
3492
Ta
b
le
4
C
on
ce
n
tr
at
io
n
ra
ti
os
fo
r
p
or
ta
bl
e
X
R
F
ve
rs
u
s
IN
A
A
.
El
Pa
so
C
or
e
M
im
br
es
-0
1
M
im
br
es
-0
2a
M
im
br
es
-0
3
M
im
br
es
-0
4a
M
im
br
es
-0
8
M
im
br
es
-1
1
M
im
br
es
-2
1
A
ll
G
ro
u
p
s
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
M
ea
n
st
d
%
rs
d
A
l
1.
03
0.
08
8
1.
13
0.
10
9
1.
13
0.
13
12
1.
31
0.
11
9
1.
06
0.
08
8
1.
03
0.
08
7
1.
34
0.
13
10
1.
23
0.
11
9
1.
16
0.
12
11
K
0.
77
0.
12
16
0.
88
0.
10
11
0.
82
0.
05
6
0.
98
0.
16
16
0.
79
0.
06
7
0.
83
0.
03
4
0.
90
0.
11
13
0.
94
0.
07
7
0.
86
0.
07
8
C
a
0.
84
0.
26
30
1.
00
0.
38
38
0.
90
0.
62
69
0.
93
0.
53
57
0.
64
0.
10
16
0.
67
0.
06
10
1.
03
0.
46
44
0.
76
0.
12
15
0.
85
0.
15
17
Ti
0.
87
0.
11
12
1.
71
0.
93
54
1.
04
0.
11
10
1.
19
0.
58
49
1.
05
0.
22
21
1.
05
0.
12
11
1.
02
0.
38
38
1.
16
0.
23
20
1.
14
0.
25
22
M
n
1.
32
0.
32
25
1.
18
0.
45
38
1.
24
0.
19
16
0.
87
0.
20
22
1.
17
0.
35
30
1.
39
0.
33
24
1.
02
0.
19
19
0.
85
0.
09
11
1.
13
0.
20
18
Fe
1.
01
0.
12
12
0.
90
0.
08
9
0.
99
0.
03
3
0.
83
0.
09
11
0.
98
0.
08
8
0.
90
0.
05
6
0.
98
0.
06
6
0.
91
0.
05
6
0.
94
0.
06
6
Th
0.
72
0.
15
20
0.
87
0.
12
14
0.
86
0.
09
10
1.
13
0.
14
12
0.
83
0.
15
18
0.
80
0.
08
10
0.
88
0.
17
19
1.
01
0.
23
23
0.
89
0.
13
14
R
b
0.
99
0.
09
9
0.
93
0.
03
3
0.
96
0.
05
5
0.
97
0.
02
2
0.
95
0.
04
4
0.
98
0.
08
8
0.
99
0.
03
3
0.
96
0.
04
4
0.
97
0.
02
2
Sr
0.
97
0.
26
27
1.
02
0.
15
14
0.
89
0.
09
11
1.
14
0.
21
18
0.
88
0.
14
16
0.
85
0.
31
37
1.
00
0.
11
11
0.
96
0.
19
20
0.
96
0.
09
10
Zr
1.
18
0.
24
20
0.
90
0.
13
15
1.
40
0.
13
10
1.
18
0.
22
18
1.
25
0.
19
15
1.
20
0.
17
14
1.
39
0.
14
10
1.
55
0.
20
13
1.
26
0.
20
16
st
d
e
st
an
d
ar
d
d
ev
ia
ti
on
.
%
rs
d
e
%
re
la
ti
ve
st
an
d
ar
d
d
ev
ia
ti
on
.
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e3496 3493linear regressions derived from the analysis of 15 obsidian refer-
ence samples (Phillips and Speakman, 2009). The obsidian cali-
bration was selected to quantify the ceramic data because our
initial experiments suggested that it worked quite well. Given that
both obsidian and pottery are silicates and given the relative
accuracy of the calibration for ceramics, this approach seemed
reasonable. Technically it also should be possible to measure zinc
using these settings; however, a high instrument background
precludes accurate quantification of this element when present in
low abundances.
For the second analysis, each sample was analyzed at 12 kV,
15 mA, for a 200-second live-time count. Peak intensities for the Ka
peaks Al, K, Ti, and Kb peak of Ca were converted to parts-per-
million (ppm) using a quadratic regression model based on the
analyses of New Ohio Red Clay (Glascock et al., 2007a,b), Brill Glass
A and B, and USGS pressed powder standards (AGV-1, DNC-1, DTS-
2, G-2, MAG-1, SDC-1, SDO-1, SGR-1, andW-2). Siliconwas observed
in all spectra and Cr was observed in some spectra, but no attempts
were made to quantify these elements.
4. Results
4.1. mXRF results
Graphical depictions of the mXRF data for 12 of the 18 elements
measured are shown in Fig. 3A and B. These line scansdbased on
measurements of Mg, Al, Si, K, Ca, Ti, Cr, Mn, Fe, Rb, Sr, and
Zrdillustrate the heterogeneous nature of the Ohio Red Clay (red
line) and archaeological pottery (blue line) relative to the more
homogeneous obsidian (black line). In the case of the obsidian
sample Cr and Mg were below detection limits. Inspection of the
line plots and the mean and standard deviations of the count rates
(see Table 2) illustrate the relative homogeneity of the obsidian
sample relative to the moderately tempered archaeological spec-
imen and the fine-paste Ohio Red Clay tile. If we exclude Mg, Cr, Ni,
Cu, Y, and Nb from consideration for all samples, due to inherently
low values in the specimens analyzed and/or instrument sensitivity
issues which resulted in greater analytical error, we see the
obsidian has an average coefficient of variation of 4% versus 19% for
the Ohio Red tile, and 35% for the archaeological sample. Visually,
significant variability in the archaeological sample is evident in the
plots for Al, K, Ca, Ti, Fe, Sr, and Zr and would seem to support
a hypothesis that mXRF is less than adequate for provenance studies
of ceramics.
4.2. Portable XRF results
Compositional data for all elements measured by portable XRF
and their corresponding INAA data are presented in Table 3.
Concentration ratios for portable XRF versus INAA groups are pre-
sented in Table 4. In general, some correlation exists between the
INAA values and the corresponding portable XRF values (Table 3).
Having said this, however, it is clear that some elements are better
correlated than others. As shown in Table 4, Rb data for portable
XRF and INAA are highly correlated and exhibit low relative stan-
dard deviations. Sr and Fe ratios for all groups are greater than 90%
and are reasonably well correlated. Mean ratios for Zr are high and
exhibit variability that likely reflects heterogeneity in the ceramic
samples measured by portable XRF and possibly higher analytical
error that is inherent in INAA measurements for this element.
Th ratios tend to be systematically low in the portable XRF data
which is a reflection of the ability of XRF to accurately measure low
ppm Th. Likewise, Al is difficult to measure by XRF and even more
challenging for samples not measured under full vacuum. We
suspect that differences among K, Ca, and Ti reflect differences
Fig. 5. Variance-covariance matrix PCA biplot based on Cr, Cs, Eu, Ta, Th measurements
by INAA that are known to be important discriminating elements of Mimbres and
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e34963494between the two type of analysesdbulk (INAA) versus heteroge-
neous surface analyses (portable XRF). Despite the observed
differences between the two analytical techniques, there is some
general agreement between the elements common to the INAA and
portable XRF analysis.
Following common statistical routines for multivariate
compositional datasets (see Glascock, 1992, among others), biplots
derived from principal components analysis (PCA) of the elements
in common to both the INAA and portable XRF datasets (Al, K, Ca,
Ti, Mn, Fe, Rb, Sr, Zr, Th), calculated here as variance-covariance
matrices, show more or less the same group structure (Fig. 4A
and B). In general, these figures illustrate the elements that can be
measured and quantified by portable XRF have similar analytical
precision and accuracy, but that these elements as a group are not
particularly useful for sourcing archaeological ceramics at a site-
specific level. When PCA scores are recalculated to include all
elements measured by portable XRF (e.g., the inclusion of Y and Nb
in the data matrix) no significant differences are observed. In
contrast, a variance-covariance matrix biplot, based on PCA of 5
INAA elements (Cr, Cs, Eu, Ta, Th) known to be good discriminatorsFig. 4. A and B Biplot derived from PCA of the variance-covariance matrix of the
Mimbres and Jornada pottery samples measured by portable XRF (A). Ellipses repre-
sent 90% confidence level for membership in the groups. Only elements in common to
both portable XRF and INAA (Al, K, Ca, Ti, Mn, Fe, Rb, Sr, Zr, Th) are used to calculate
PCA scores. (B) Same as above, but based on INAA data.
Jornada pottery.
Fig. 6. A and B Elemental plots showing the best separation of groups/samples
analyzed by portable XRF (A) and INAA (B).
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e3496 3495of the groups in question, show a clear and unambiguous separa-
tion of the eight compositional groups (Fig. 5).
In addition to examining the data in multivariate space,
attempts were made to show group discrimination using bivariate
plots of the elements. The absolute best possible separation of the 8
groups using portable XRF data is shown in Fig. 6A. In this figure Th
and Nb show marginal separation of the El Paso Core and groups
Mimbres-1, Mimbres-3, and Mimbres-21 from the 4 Mimbres
Valley groups (Mimbres-2a, 4a, 8, and 11). The four Mimbres Valley
groups can be separated using Sr and Rb (Fig. 6A, inset). In contrast
to the portable XRF data, excellent separation of El Paso Core and
groups 1, 3, 8, and 21 is observed in a bivariate plot based on Th and
Ta INAA concentrations (Fig. 6B). The three remaining Mimbres
Valley groups exhibit equally good discrimination using Cs and Cr
(Fig. 6B, inset).
5. Conclusions
Portable XRF has demonstrated great potential for quantitative
analyses of manymaterial classes, such as obsidian andmetals (e.g.,
Heginbotham et al., 2010; Phillips and Speakman, 2009); however,
the application of portable XRF to provenance studies of ceramics is
not straightforward. In this study we examined 75 Mimbres and
Jornada pottery samples by portable XRF.We carefully prepared the
surfaces of the sherds and made every effort to optimize the
portable XRF experiment to ensure that we generated the best
possible data. The results of the portable XRF analyses were directly
compared to INAA data generated from the same sherds. Although
there is not a direct 1:1 correlation between the two datasets, data
from the portable XRF study are somewhat in agreement with the
corresponding INAA data. In particular, portable XRF was found to
be suitable for distinguishing between major drainage basins and,
to some extent, even short-distance intradrainage variation as
shown by our results. Despite the generally acceptable results from
portable XRF, it is clear that INAA has significantly greater analytical
precision. In addition, the ability of INAA to measure trace and rare
earth elements has proved critical to being able to effectively
identify compositional groups that are useful for provenance
studies. We suspect, based on our knowledge of other studies in the
Americas, that the same can be said for most pottery studiesdthat
is to say that analytical precision and the ability to measure trace
and rare earth elements is key for discriminating compositional
groups at a level required for site-specific (or approximately so)
research questions. We are certain that there are some exemptions
to this rule, but would argue that a portable-XRF study of ceramics
should never be undertaken blindly. It is important to have some
priori knowledge of the chemical variability and the expected group
structure so that meaningful compositional groups can be
constructed.
In cases where museums will not permit removal of samples
from whole pottery vessels (often referred to as destructive
sampling), thus precluding INAA and/or ICP-MS, portable XRF may
be the only practical method of chemical compositional analysis
currently available. In such cases, some data are better than none
even if the results may be less useful for answering some research
questions than INAA and/or ICP-MS.
As a final consideration, the mXRF experiments we conducted
highlighted significant chemical variability within individual
ceramics samples. The heterogeneous nature of ceramics at the mm
scale must be considered when initiating provenance studies by
mXRF. There is a big difference between the 80 mm diameter spot
used in our mXRF experiments and the ca. 4000 mmdiameter spot of
our portable XRF instrument. Although much of our ability to
generate numbers comparable to INAA lies in the preparation of our
samples, we must also consider that larger beam sizes will averageout some of the spatial chemical variability that occurs in ceramic
pastes. It is therefore important to underscore that the ability to
generate quantitative mXRF data for archaeological ceramics will
be challenging and likely will have limited applicability to ceramic
provenance studies.
Acknowledgments
We acknowledge all the Southwestern archaeologists who have
submitted Mimbres, Jornada, and related ceramics for INAA at
Texas A&M and MURR. Steve Shackley, Mike Glascock, Steven
LeBlanc, and three anonymous reviewers are thanked for their
thoughtful comments and criticisms of an earlier draft of this paper.
INAA analyses were conducted at the University of Missouri
Research Reactor and subsidized by various NSF grants awarded to
Glascock and colleagues. JGI is indebted to the support of the Marie
Curie International Outgoing Fellowships program, endorsed by
the European Commission (“ARCHSYMB”, PIOF-GA-2008-223319).
References
Aldenderfer, M., Craig, N.M., Speakman, R.J., Popelka-Filcoff, R., 2008. Four-thou-
sand-year-old gold artifacts from the Lake Titicaca Basin, Southern Peru.
Proceedings of the National Academy of Sciences of the United States of
America 105, 5002e5005.
Bishop, R.L., Rands, R.L., Holley, G.R., 1982. Ceramic compositional analysis in
archaeological perspective. Advances in Archaeological Method and Theory 5,
275e330.
Brostoff, L.B., Maynor, C., Speakman, R.J., 2009. Preliminary study of a Georgia
O’Keeffe pastel drawing by XRF and mXRD. Powder Diffraction 24, 116e123.
Buxeda, J., Cau Ontiveros, M.A., Kilikoglou, V., 2003. Chemical variability in clays
and pottery from a traditional cooking pot production village: testing
assumptions in Pereruela. Archaeometry 45, 1e17.
Cecil, L.G., Moriarty, M.D., Speakman, R.J., Glascock, M.D., 2007. Feasibility of field-
portable XRF to identify obsidian sources in Central Petén, Guatemala. In:
Glascock, M.D., Speakman, R.J., Popelka-Filcoff, R.S. (Eds.), Archaeological
Chemistry: Analytical Methods and Archaeological Interpretation. ACS Publi-
cation Series, vol. 968. American Chemical Society, Washington, DC,
pp. 506e521.
Cesareo, R., Sciuti, S., Marabelli, M., 1973. Non-destructive analysis of ancient
bronzes. Studies in Conservation 18 (2), 64e80.
Charlton, M., Crew, P., Rehren, Th., Shennan, S., 2010. Explaining the evolution of
ironmaking recipesdan example from northwest Wales. Journal of Anthropo-
logical Archaeology 29, 352e367.
Cogswell, J., Neff, H., Glascock, M., 1996. The effect of firing temperature on the
elemental characterization of pottery. Journal of Archaeological Science 23,
283e287.
Craig, N., Speakman, R.J., Popelka-Filcoff, R.S., Aldenderfer, M., Flores Blanco, L.,
Vega, M.B., Glascock, M.D., Stanish, C., 2010. Identifying the location of the
Macusanite obsidian source in southern Peru. Journal of Archaeological Science
37, 569e576.
Craig, N., Speakman, R.J., Popelka-Filcoff, R.S., Glascock, M.D., Robertson, J.D.,
Shackley, M.S., Aldenderfer, M.S., 2007. Comparison of XRF and PXRF for anal-
ysis of archaeological obsidian from southern Peru. Journal of Archaeological
Science 34, 2012e2024.
Emery, V.L., Morgenstein, M., 2007. Portable EDXRF analysis of a mud brick
necropolis enclosure: evidence of work organization, El Hibeh, Middle Egypt.
Journal of Archaeological Science 34, 111e122.
Farris, D.W., Jaramillo, C., Bayona, G., Restrepo, S.A., Montes, C., Cardona, A., Mora, A.,
Speakman, R.J., Glascock, M.D., Reiners, P., Valencia, V., Fracturing of the Pan-
amanian Isthmus during initial collision with South America. Geology, in press.
Gilman, P.A., Canouts, V., Bishop, R.L., 1994. The production and distribution of
classic Mimbres black-on-white pottery. American Antiquity 59, 695e709.
Glascock, M.D., 1992. Characterization of archaeological ceramics at MURR by
neutron activation analysis and multivariate statistics. In: Neff, H. (Ed.),
Chemical Characterization of Ceramic Pastes in Archaeology. Prehistory Press,
Madison, Wisconsin, pp. 11e26.
Glascock, M.D., Speakman, R.J., Burger, R.L., 2007a. Chemical characterization of
archaeologically important obsidian sources in Peru. In: Glascock, M.D.,
Speakman, R.J., Popelka-Filcoff, R.S. (Eds.), Archaeological Chemistry: Analytical
Methods and Archaeological Interpretation. ACS Publication Series, vol. 968.
American Chemical Society, Washington, DC, pp. 522e544.
Glascock, M.D., Speakman, R.J., Neff, H., 2007b. Archaeometry at the University of
Missouri research reactor and the provenance of obsidian artifacts in eastern
north America. Archaeometry 49, 343e357.
Goebel, T., Speakman, R.J., Reuther, J.D., 2008. Results of geochemical analysis of
obsidian artifacts from the Walker Road Site, Alaska. Current Research in the
Pleistocene 25, 88e90.
R.J. Speakman et al. / Journal of Archaeological Science 38 (2011) 3483e34963496Goren, Y., Mommsen, H., Klinger, J., 2011. Non-destructive provenance study of
cuneiform tablets using portable X-ray fluorescence (pXRF). Journal of
Archaeological Science 38, 684e696.
Grissom, C., Gervais, C., Little, N.C., Bieniosek, G., Speakman, R.J., 2010. Red "stain-
ing" on marble: chemical or biological origin. APT Bulletin: Journal of Preser-
vation Technology 41 (2e3), 11e20.
Guerra, M.F., 1998. Analysis of archaeological metals. The place of XRF and PIXE in
the determination of technology and provenance. X-Ray Spectrometry 27,
73e80.
Hall, M.E., 2004. Pottery production during the Late Jomon period: insights from the
chemical analyses of Kasori B pottery. Journal of Archaeological Science 31,
1439e1450.
Heginbotham, A., Bezur, A., Bouchard, M., Davis, J., Eremin, K., Frantz, J.H.,
Glinsman, L., Hayek, L., Hook, D., Kantarelou, V., Karydas, A., Lee, L., Mass, J.,
Matsen, C., McCarthy, B., McGath, B., Shugar, A., Sirois, J., Smith, D.,
Speakman, R., 2010. A round-robin study to evaluate the inter-laboratory
reproducibility of quantitative X-ray fluorescence spectroscopy on historic
copper alloys. In: Mardikian, P., Chemello, C., Watters, C., Hull, P. (Eds.), METAL
2010 (CD). Clemson University, South Carolina, pp. 178e188.
Hein, A., Tsolakidou, A., Iliopoulos, I., Mommsen, H., Garrigós, J.Buxeda I.,
Montana, G., Kilikoglou, V., 2002. Standardisation of elemental analytical
techniques applied to provenance studies of archaeological ceramics: an inter
laboratory calibration study. Analyst 127, 542e553.
Hughes, R.E., Högberg, A., Olausson, D., 2010. Sourcing flint from Sweden and
Denmark: a pilot study employing non-destructive energy dispersive X-ray
fluorescence spectrometry. Journal of Nordic Archaeological Science 17, 15e25.
Iñañez, J.G., Garrigós, J. Buxeda I., Speakman, R.J., Glascock, M.D., Sosa Suárez, E.,
2007. Characterization of 15the16th Century majolica pottery found on the
Canary Islands. In: Glascock, M.D., Speakman, R.J., Popelka-Filcoff, R.S. (Eds.),
Archaeological Chemistry: Analytical Techniques and Archaeological Interpre-
tation. American Chemical Society, Washington D.C., pp. 376e398.
Little, N.C., Florey, V., Molina, I., Owsley, D.W., Speakman, R.J., Measuring heavy
metal content in bone using portable XRF. Open Journal of Archaeometry, in
press.
Marwick, B., 2005. Element concentrations and magnetic susceptibility of anthro-
sols: indicators of prehistoric human occupation in the inland Pilbara, Western
Australia. Journal of Archaeological Science 32, 1357e1368.
Morgenstein, M., Redmount, C.A., 2005. Using portable energy dispersive X-ray
fluorescence (EDXRF) analysis for on-site study of ceramic sherds at El Hibeh,
Egypt. Journal of Archaeological Science 32, 1613e1623.
Neff, H., 2000. Neutron activation analysis for provenance determination in
archaeology. In: Ciliberto, E., Spoto, G. (Eds.), Modern Analytical Methods in Art
and Archaeology. Wiley, New York, pp. 81e134.
Pantazis, T., Karydas, A.G., Doumas, C., Vlachopoulos, A., Nomikos, P., Dinsmore, M.,
2002. X-ray Fluorescence Analysis of a Gold Ibex and Other Artifacts from
Akrotiri Paper presented at the 9th International Aegean Conference at Yale
University, April 18e21, 2002.
Phillips, S.C., Speakman, R.J., 2009. Initial source evaluation of archaeological
obsidian from the Kuril Islands of the Russian Far East by portable XRF. Journal
of Archaeological Science 36, 1256e1263.Piorek, S., 1997. Field-portable X-ray fluorescence spectrometry: past, present, and
future. Field Analytical Chemistry and Technology 1, 317e329.
Potts, P.J., Webb, P.C., Williams-Thorpe, O., 1995. Analysis of silicate rocks using
field-portable X-ray fluorescence instrumentation incorporating a mercury (II)
iodide detector: a preliminary assessment of analytical performance. The
Analyst 120, 1273e1278.
Potts, P.J., Williams-Thorpe, O., Webb, P.C., 1997a. the bulk analysis of silicate rocks
by portable X-ray fluorescence: effect of sample mineralogy in relation to the
size of the excited volume. Geostandards Newsletter 21 (1), 29e41.
Potts, P.J., Webb, P.C., Williams-Thorpe, O., 1997b. Investigation of a correction
procedure for surface irregularity effects based on scatter peak intensities in the
field analysis of geological and archaeological rock samples by portable X-ray
fluorescence spectrometry. Journal of Analytical Atomic Spectrometry 12,
769e776.
Reuther, J.D., Slobodina, N.S., Rasic, J.T., Cook, J.P., Speakman, R.J., 2011. Gaining
momentumdthe status of obsidian source studies in Alaska: implications for
understanding regional prehistory. In: Goebel, T., Buvit, I. (Eds.), From the
Yenisei to the Yukon Interpreting Lithic Assemblage Variability in Late Pleis-
tocene/Early Holocene Beringia. Texas A&M University Press, College Station,
pp. 270e286.
Shackley, M.S., 1998. Geochemical differentiation and prehistoric procurement of
obsidian in the Mount Taylor volcanic field, northwest New Mexico. Journal of
Archaeological Science 25, 1073e1082.
Shackley, M.S., 2010a. Is there reliability and validity in portable X-ray fluorescence
spectrometry. The SAA Archaeological Record 10 (5), 17e20. 44.
Shackley, M.S., 2010b. An introduction to X-ray fluorescence (XRF) analysis in
archaeology. In: Shackley, M.S. (Ed.), X-ray Fluorescence Spectrometry (XRF) in
Geoarchaeology. Springer, New York, pp. 7e44.
Slobodina, N.S., Reuther, J.D., Rasic, J.T., Cook, J.P., Speakman, R.J., 2009. Obsidian
procurement and use at the dry Creek site (HEA-005), Interior Alaska. Current
Research in the Pleistocene 26, 115e117.
Speakman, R.J., Holmes, C.E., Glascock, M.D., 2007. Source determination of obsidian
artifacts from Swan Point (XBD-156), Alaska. Current Research in the Pleisto-
cene 24, 143e145.
Tagle, R., Gross, A., 2010. ARTAX/TRACER III-V Provenance Studies on Ceramics with
Portable (m)XRF Spectrometers Lab Report XRF 437, Bruker Nano GmbH, Berlin.
Weigand, P.C., Harbottle, G., Sayre, E.V., 1977. Turquoise sources and source analysis:
Mesoamerica and the Southwestern U.S.A. In: Earle, T.K., Ericson, J.E. (Eds.),
Exchange Systems in Prehistory. Academic Press, New York, pp. 15e34.
Williams-Thorpe, O., Webb, P.C., Jones, M.C., 2003. Non-destructive geochemical
and magnetic characterisation of Group XVIII dolerite stone axes and shaft-hole
implements from England. Journal of Archaeological Science 30, 1237e1267.
Williams-Thorpe, O., Potts, P.J., Webb, P.C., 1999. Field-portable non-destructive
analysis of lithic archaeological samples by X-ray fluorescence instrumenta-
tion using a mercury iodide detector: comparison with wavelength-dispersive
XRF and a case study in British stone axe provenancing. Journal of Archaeo-
logical Science 26, 215e237.
Wolff, C., Fitzhugh, W., Speakman, R.J., The utility of pXRF in the assessment of slate
procurement and exchange by the Maritime Archaic of Newfoundland and
Labrador. Open Journal of Archaeometry, in press.